https://github.com/alexis-jacq/LOLA_DiCE

Pytorch implementation of LOLA (https://arxiv.org/abs/1709.04326) using DiCE (https://arxiv.org/abs/1802.05098)
https://github.com/alexis-jacq/LOLA_DiCE

Last synced: 2 months ago
JSON representation

Pytorch implementation of LOLA (https://arxiv.org/abs/1709.04326) using DiCE (https://arxiv.org/abs/1802.05098)

• Host: GitHub
• URL: https://github.com/alexis-jacq/LOLA_DiCE
• Owner: alexis-jacq
• License: mit
• Created: 2018-07-31T12:09:50.000Z (almost 6 years ago)
• Default Branch: master
• Last Pushed: 2018-08-21T03:02:05.000Z (almost 6 years ago)
• Last Synced: 2024-01-26T20:34:50.426Z (5 months ago)
• Language: Python
• Size: 285 KB
• Stars: 88
• Watchers: 6
• Forks: 15
• Open Issues: 4
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Lists

• Awesome-pytorch-list - LOLA_DiCE
• Awesome-pytorch-list-CNVersion - LOLA_DiCE

README

        
# LOLA_DiCE

Pytorch implementation of LOLA (https://arxiv.org/abs/1709.04326) using DiCE (https://arxiv.org/abs/1802.05098)

## Quick results:

### Results on IPD using DiCE 

[lr_in=0.3, lr_out=0.2, lr_v=0.1, batch_size=128, len_rollout=150, use_baseline=True]

![ipd_with_dice](images/ipd_dice.png)

### Results on IPD using DiCE and opponent modelling

[lr_in=0.3, lr_out=0.2, lr_v=0.1, batch_size=128, len_rollout=150, use_baseline=True]

![ipd_with_dice](images/ipd_dice_om.png)

(It seems that 2 lookaheads is the most stable model with this set of hyperparameters)

### Results on IPD using exact gradients 

[lr_in=0.3, lr_out=0.2, batch_size=128, len_rollout=150]

![ipd_with_exact_grads](images/ipd_exact.png)

### Results on IPD using exact gradients and opponent modelling

[lr_in=0.3, lr_out=0.2, batch_size=128, len_rollout=150]

![ipd_with_exact_grads](images/ipd_exact_om.png)

## Authors version:

The authors of the paper have their own version (Tensorflow) available here: https://github.com/alshedivat/lola